Based on your input and expected result, I would guess that you are looking to regrid your data from the x/y curvilinear grid of your GRIB weather data to a new rectilinear (lon/lat) grid?

If so, I would suggest using xESMF (example here), since xarray itself does not appear to support this type of interpolation yet. As an example, given the dimension names you currently have and expect in the output, your code could look something like the following:

```python import xesmf as xe

...load data...

ds_xy_grid = ds_xy_grid.rename(latitude='lat', longitude='lon') ds_out = xr.Dataset({'lat': (['lat'], np.linspace(44.77, 56.14, 421)), 'lon': (['lon'], np.linspace(2.976, 19.84, 461)}) regridder = xe.Regridder(ds_xy_grid, ds_out, 'bilinear') ds_lonlat_grid = regridder(ds_xy_grid) ds_lonlat_grid = ds_lonlat_grid.rename(lat='latitude', lon='longitude') ```

(updated as per comment below)